Interpretable Process Mining: Shifting Control to End Users

Professor Marcello La Rosa

School of Computing and Information Systems

University of Melbourne

Abstract: Process mining is an innovative technology for extracting process insights from transactional logs recorded by common IT systems, in order to analyse and improve organisational productivity along performance dimensions such as efficiency, quality, compliance and risk. By using data rather than user perceptions, process mining offers an unbiased view on current and past process performance, effectively promoting a paradigm shift from “confidence-based” to “evidence-based” business process management. But, are we ready to realise the process mining promise of rapid and effective process management? While to date many organisations have employed process mining with various degrees of success, there are still open challenges that prevent its widespread adoption in practice. One interesting challenge relates to “interpretability”: from an end-user perspective, how easy is it to interpret and act upon the insights offered by process mining? This presentation will illustrate several approaches aimed at lifting the interpretability of the full spectrum of process mining technology.

Bio: Professor Marcello La Rosa leads the Information Systems group within The University of Melbourne’s School of Computing and Information Systems. Before that he was a Professor with the Queensland University of Technology, where he led the Business Process Management (BPM) discipline and served as the director for corporate programs and partnerships. He was also the recipient of an Information Systems Fellowship from the University of Liechtenstein and held a part-time Principal Researcher position at NICTA (now Data61). His research interests span different BPM areas with a focus on process mining, process consolidation and automation, in which he published extensively. He leads the Apromore initiative, a strategic inter-university collaboration for the development of an open-source process analytics platform. Marcello has taught BPM to practitioners and students in Australia and overseas for over ten years. Based on this experience, he co-authored the first, comprehensive textbook on BPM (Springer, 2nd edition), which has influenced the curriculum of over 200 universities in the world. Using this book, he co-developed a series of MOOCs on the subject, which have attracted over 25,000 participants to date.